Edge Detection Algorithm Based On Lifting Wavelet An Morphology For Medical Image

Medical imaging has become an important basis and mean for clinical diagnosis, pathological analysis and treatment. Medical image processing provides a great help for medical research and clinical diagnosis and treatment. Due to the medical imaging showing a very rich inside information of the body, the body’s internal lesions can be showed in a more intuitive and clear way. Edge detection is a vital part in medical image processing, and whether the medical image edge detection good or bad will directly affect the subsequent treatment.Medical image edge detection is a process that extracting the lesion part’s boundary from medical images. Now several commonly used edge detection methods mostly use local image differential to extract image edge, such as gradient operator, Laplace-Gaussian operator. However using differential to extract edge is often very sensitive to noise, so that the extracted edges become blurred.Edge detection methods based on lifting wavelet and mathematical morphology are emerging in recent years. Not only inherits the lifting wavelet multi-scale analysis capability and good time-frequency localization characteristics from traditional wavelet, but also has the advantage of fast calculation speed and less memory occupation, which is very suitable for image processing. Mathematical morphological method uses fixed structural elements and morphological operators to measure and recognize the shapes in the images. Thereby mathematical morphology edge detection algorithm can preserve more image detail information and remove the noise effectively, and also has the advantage of simple arithmetic、parallel processing and easy hardware implementation.This paper will combine the advantages of lifting wavelet transform and mathematical morphology, change them and propose new edge detection algorithm for medical image. This new edge detection algorithm will be able to effectively extract relatively complete and accurate positioning medical image edge. The main work of this paper is as follows:(1)Edge detection is implemented by using the classical differential edge detection operators including Robert operator, Sobel operator, Prewitt operator, Laplace operator and so on for medical images one by one, and analyze and compare their advantages and disadvantages.(2)The basic theory of lifting wavelet analysis and mathematical morphology is discussed, and the experimental results and analysis of their edge detection are given.(3)Combined with lifting wavelet and morphology, three new edge detection algorithms are proposed, and through a large number of experiments good experimental results have been got. The classic edge detection algorithms were compared and analyzed. These algorithms are also used on noised medical image for edge detection, and image edge detection analysis results show that the methods have good positioning accuracy and anti-noise ability.